Surface-Reflected GPS Wind Speed Sensing Results for 2010 Atlantic Season

Surface-Reflected GPS Wind Speed Sensing Results for 2010 Atlantic Season

6565thth Interdepartmental Hurricane Conference Interdepartmental Hurricane ConferenceMiami, FL

Feb. 28 – March 3, 2011

Michael S. Grant, (NASA/Langley Research Center)

Stephen J. Katzberg,(NASA/Distinguished Research Associate)

Jason P. Dunion (Univ. Miami/NOAA/AOML/Hurricane Research Division)

Surface-Reflected (Bistatic) GPS Method

2010 Storm Season Wind Speed Retrievals Retrieval examples Quantitative comparisons to SFMR and dropsondes

Summary Statistics for Measurement Comparisons

Future Research Objectives

Presentation Outline

Ocean Roughness / Wind Speed from Bistatic Surface Reflections

Constant Path Delay Ellipses in Reflection Area

z

Py

h

x

Reflection Area

GPS GPS

Sγ

Increasing Surface Roughness

(Instrument Correlation Peak)

delay, τ

Cor

rela

tion

(Ref

lect

ed P

ower

)

Ocean ‘roughness’ (surface slope variance) used to infer surface wind speed Slope-to-wind speed: empirical relationship.

Reflected GPS signal strength (~ power) vs. delay is measured Waveform widens (more scattering) with increasing surface roughness. Sensing location on surface depends on satellite-aircraft reflection path geometry

GPS Instrument easily deployed Light aircraft and up Flown (2004) Aerosonde UAV – 10’ wing span

Instrument - Receiver Unit and two Antennae

Instrument size ~ 16 x 12 x 7 inches Weight < 10 lbs. 3.5” nadir antenna

NASA Bistatic GPS Instrument Accommodation

NASA-Langley Bistatic GPS Instrument

AOC WP-3D Orion

Cessna 206

Aerosonde UAV

NASA-Langley bistatic GPS instruments deployed on ‘N42 and ‘N43 P-3 Hurricane Hunter Aircraft (Aircraft Operations Center, Tampa)

Instruments operated by AOC personnel (Power on, autonomous operation, power off, upload flight data to ftp site post-mission)

31 total P-3 flights (incl. ferry) where GPS instruments were operated. Data sets acquired on all flights – no instrument anomalies. 18 GPS data sets had contemporaneous data available for comparison:

SFMR, Flight-Level winds, and dropsondes (16 of 18) Variety of SFMR/dropsonde wind speeds ranges for the 18 GPS data sets:

lowest: 2 – 12 ms-1

highest: 5 – 60+ ms-1

For 2010 quantitative comparisons, only GPS over-land reflections data removed. No other data exclusions or masking operations were performed.

GPS Instrument 2010 Atlantic Storm Season Deployments

Stepped-Frequency Microwave Radiometer (SFMR) [1]

Operational instrument, high precision (within 2% at 30 ms-1) [2]

Emissivity (brightness temp., TB), surface wind speed proportional to % sea foam.

NASA-Langley Bistatic GPS Instrument - surface wind speed obtained from measured sea-surface slopes through empirical relationship [3][4]

[1] Black, P. G., and C. L. Swift, 1984: Airborne stepped frequency microwave radiometer measurements of rainfall rate and surface wind speed in hurricanes. Second Conf. on Radar Meteorology, Zurich, Switzerland, Amer. Meteor. Soc.

[2] Uhlhorn, Black, et al., Hurricane Surface Wind Measurements from an Operational Stepped Frequency Microwave Radiometer, Monthly Weather Review, 2007, Vol. 135, p. 3070

Wind Speed Retrieval Intercomparisons - Background

[3] Katzberg, Stephen J., Omar Torres, and George Ganoe, “Calibration of reflected GPS for tropical storm wind speed retrievals”; Geophys. Res. Lett., 33, L18602, doi:10.1029/2006GL026825, 2006[4] Katzberg, S. J., and J. Dunion , “Comparison of reflected GPS wind speed retrievals with dropsondes in tropical cyclones,” Geophys. Res. Lett., 36, L17602, doi:10.1029/2009GL039512., 2009

Initial calibration using Navy COAMPS model [3]

Comparison with dropsondes at hurricane wind speeds [4]

Low-wind speed example -Tropical Depression #5 (Aug 11, 2010)

Wind Speed Retrieval Intercomparisons - Quantitative

7.2 7.4 7.6 7.8 8 8.2 8.4 8.6 8.8 9 9.2

x 104

0

5

10

15

20

25

30

35

40

45

UTC "today"(sec)

U10

WS

pd (m

/s)

GPS-Derived Surface Winds - Trop Depression #5, N42 (Aug 11, 2010)

GPS 2s-AvgSFMR0.8Flt-LvlSondes

Histogram of (SFMR – GPS) differences (ms-1)

99oW 90oW

81o W

72o W

63o W

16o N

24o N

32o N

40o N

Begin EndLon

Lat

Flight Path (GPS) - Trop Depression #5, N42 (Aug 11, 2010)

Comparison to SFMR Mean diff. = 4.3 ms-1 (GPS underestimate) RMS diff. = 2.3 ms-1

Comparison to dropsondes Mean diff. = -0.5 ms-1 (GPS overestimate) RMS diff. = 2.9 ms-1

High-wind speed example 1 – Hurricane Earl (Sept 1, 2010)

Wind Speed Retrieval Intercomparisons – Quantitative, cont’d

Histogram of (SFMR – GPS) differences (ms-1)

Comparison to SFMR Mean diff. = -0.4 ms-1

RMS diff. = 7.2 ms-1

12 - 20 ms-1 difference at eyewall

Comparison to dropsondes Mean diff. = 2.7 ms-1 (HSA-file, single near-surface value)

8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6

x 104

0

10

20

30

40

50

60

70

UTC "today"(sec)

U10

WS

pd (m

/s)

GPS-Derived Surface Winds - Hurr. Earl, N43 (Sept 01, 2010) (>30 sat-elev)

GPS mdl-filtSFMR0.8Flt-LvlSondes

96oW 84

o W 72o W 60

o W

48o W

0o

12o N

24o N

36o N

48o N

Begin

End

Lon

Lat

Flight Path (GPS) - N43, Sept 01, 2010

High-wind speed example 2 – Hurricane Earl along East Coast (Sept 3, 2010)

Wind Speed Retrieval Intercomparisons – Quantitative, cont’d

Comparison to SFMR Mean diff. = 5.9 ms-1

RMS diff. = 7.2 ms-1

Comparison to dropsondes Mean diff. = 4.2 ms-1 (GPS underestimate) RMS diff. = 7.2 ms-1

Histogram of (SFMR – GPS) differences (ms-1)

90o W

80o W 7

0o W

60o W

50o W

18o N

27o N

36

o N

45o N

54o N

Begin

End

Lon

Lat

Flight Path (GPS) - N42, Sept 03, 2010

4.2 4.4 4.6 4.8 5 5.2 5.4 5.6 5.8

x 104

0

10

20

30

40

50

60

70

UTC "today"(sec)

U10

WS

pd (m

/s)

GPS-Derived Surface Winds - Hurr. Earl, N42 (Sept 03, 2010) (>30 sat-elev)

GPS mdl-filtSFMR0.8Flt-LvlSondes

GPS measurement differences with dropsondes and SFMR generally increased with storm maximum wind speed.

Wind Speed Retrieval Intercomparisons – Summary Statistics

10 15 20 25 30 35 40 45 50 55 60 65

2

4

6

8

10

12

14

16

Max Wspd (ms-1)

RM

S D

ev. (

ms-1

)

RMS Deviations Relative to SFMR vs. Max Wind Speed

data 1 linear

Higher variability and larger range in true wind speed.

10 15 20 25 30 35 40 45 50 55 60 65 70-10

-5

0

5

10

15

20

Max Wspd (ms-1)

Mea

n D

ev. (

ms-1

)

Mean Deviations Relative to SFMR and Dropsondes vs. Max Wind Speed

(Rel. to SFMR)(Rel. to Drops.) Primarily due

to hurricane eyewall.

GPS wspd. absolute mean differences: 0.3 – 7.5 ms-1 (SFMR) 0.1 – 16.3 ms-1 (drops.)

GPS wspd. RMS differences:

2.3 – 11.8 ms-1 (SFMR) 0.4 – 8.9 ms-1 (drops.)

The 18 retrievals (data sets) were categorized in 3 maximum wind speed ranges ( 6 sets per category )

Wind Speed Retrieval Intercomparisons – Summary Statistics, cont’d

Max Wspd

Aggregate GPS Measurement Performance per Max Wind Speed Category (all entries in ms-1)

12 - 20

(vs. SFMR)

28 - 35

40 - 62

0.6 ± 3.6

3.7 ± 9.3

0.4 ± 11.8

(vs. Dropsondes)

Category Differences: Mean ± 1σ †

1.5 ± 4.0

2.7 ± 7.9

7.8 ± 8.9

† σ is the largest RMS diff. in each category

NASA Bistatic GPS instruments on NOAA ‘N42 and ‘N43 a/crft performed well No anomalies, equivalent retrieval quality from each instr.

GPS-derived surface wind speed generally compared well to SFMR and dropsonde measurements.

Best performance (vs. dropsondes) currently over the 0 – 35 ms-1 range: bias (underestimate) less than 3 ms-1 precision better than 4 ms-1 (1σ): 0 – 20 ms-1 precision better than 8 ms-1 (1σ): 0 – 35 ms-1

Measurements with peak winds in 40 – 60+ ms-1 range: bias (underestimate) less than 8 ms-1

precision better than 9 ms-1 (1σ) Significant underestimates primarily of winds in hurricane eyewall

Future Reduce bias in peak wind/eyewall measurements and improve precision for P-3

hurricane/TC missions. GPS wind speed assimilation in intensity forecasting (or other) models

Data product quality control and accuracy, precision reqt’s

Add missions/platforms for wind field mapping of developing storm systems.

Summary, Future Objectives

Acknowledgements

~ For the excellent support ~Thank you !!

Dr. James McFadden Dr. James McFadden (AOC Chief, Programs and Projects)(AOC Chief, Programs and Projects)

Terry Lynch Terry Lynch (Chief, Technical Section)(Chief, Technical Section)

Joe Bosko Joe Bosko (N42RF Science systems)(N42RF Science systems)

Dana NaeherDana Naeher (N43RF Science systems(N43RF Science systems))